Thursday, May 10, 2012

Fall of France, VI: Traction Avant!

We live in a world on the verge of chronic underemployment and incipient deflationary spiral. I'd say that it's hard to understand how this came about, except that I'm an early modern historian, so I've seen it before. People who have ready money appreciate perfectly well that it is most valuable when no-one else has it, and people with ready money can usually get the prince to go along with their desires. There'd be no economic growth at all were it not for that great social wonder called "war," which (sometimes) forces the rich to confront the alternative of losing office with the fall of the dynasty. Nothing opens the hoards and sets money into circulation like the thought of your parish rival getting your family's J.P. by coming out for Charlie in time!

Sometimes. Not always. The last century saw two great wars. One was followed by deflation, the other by what I'm increasingly inclined to see as the single greatest economic expansion in human history. If we could just find a way to bottle that stuff, maybe I could afford a house in my neighbourhood. We just don't want to bottle WWI when we mean to be bottling WWII, New Coke instead of Old Coke. So what makes the difference? For one, France won in 1914, lost in 1940. Is there anything, anything, we can learn from that?

How about that Citroën drivers are a weird cult? 

There's older, weirder video footage in the long history of M. Citroën's long history, I'm told. I certainly find "weirder" on Youtube. It's the "older" that's beating me. The top Youtube search result for "Traction Avant" is a 55 second video of somebody's dashboard taken as he drives through Vienna. I made one of those on the Island Highway the other day, but at least I'm willing to admit that I was trying to take the seascape, and delete it, as opposed to posting it.

Okay, wait, Google turned up the "Traction Avant" song. You're welcome. And in the spirit of acknowledging brave efforts, here's Alessandro Galati trying to reclaim the tag from the ashbin of history.

My point, such as it is, is this: At one level, the Allies' defeat in 1940 is not very complicated. Germany had air superiority, and this allowed German bombers to disrupt the French artillery fireplan for the defence of the Sedan bend of the Meuse. Obviously, at another level, it is. I've talked about why the French artillery plan could be so disrupted, and why the BEF was in no position to counterattack, and, indeed, tied up French counterattack assets.

What I haven't talked about is the reason that the Armee de l'Air lost air superiority. It's a story that's usually  accounted for with production statistics, usually accompanied with snide allegations that communistic French workers sabotaged all the planes. It's certainly true that France wasn't building enough aeroplanes in 1938. But it's not the whole story. This needs to be a story about industry,  and what might be different as between  1914 and 1940.

This is not a subject that the French like to explore, even in the tongue-to-sore-tooth sort of way. But there was that classic number of Revues Guerres Mondiale that published R. Danel and Ingenieur-General Truel's research.  Per Danel, we know that on 16 August, 1939, the French possessed 7,450 a/c in total, not counting 353 a/c of the navy. These were divided into: 3,959 warplanes, 2691 trainers and other auxiliaries such as air ambulances, and 800 “avions prémilitaires (aviation populaire)”. 

Of the military a/c, 1264 old or very old aircraft included 813 in North Africa, plus squadrons, schools, trial centres, training centres, and storage centres, which issued 160 to the “colonies.” An additional 1617 a/c were transitional types, of which 1377 were in service (plus 47 in the colonies; and 1078 were modern types, 752 in service, plus 55 in the colonies. In total, the formations in the metropole and North Africa amounted to 2942 warplanes ancient and modern, of which 752 were in storage, and 262 were in the colonies.
Older Aircraft (prior to Plans I and II)
 Fighters: 135 Nieuport Delage 62 and 622, 18 Nieuport-Delage 629, 21 Morane Saulnier 225, in total 174 fighters. Bombers: 102 Loiré-et-Olivier 20 and 29 Loiré-et-Olivier 206, a total of 131. Observation: 61 Potez 39 and 96 Bréguet 27, a total of 157. “Divers:” 475 Potez 25 métropole and 327 Potez 25 “T.O.E.” a total of 802.
Transition Aircraft (plans I and II)
Fighter: 16 Dewoitine 371, 67 Dewoitine 500, 107 Dewoitine 501, 81 Dewoitine 510, 51 Loire 46, 49 Spad 510, and 28 Mureaux 113 CN2, a total of 399. Bomb and recon: 126 Amiot 143, 169 Bloch 200, 238 Bloch 210, 42 Farman 221 and 222, Loiré-et-Olivier 257bis, 172 Potez 540, 60 Potez 542 and 124 Bloch 131, a total of 950. Observation: 100 Les Mureaux 115, 93 Les Mureaux 117, and 52 Loiré-et-Olivier autogiros, a total of 245.

More importantly, the modern aircraft of Plan V included: SS Fighters, 88 Bloch 151 and 152, 152 Curtiss H75 A1 and A2; Multiplace fighters, day/night: 277 Potez 630 and 631; Bombers, 15 Potez 633, 4 Loiré-et-Olivier 451 and 1 Loire-Nieuport 411, a total of 20 aircraft. Reconaissance: 59 Potez 637. Troop transports: 6 Farman 224.
“L’Armée de l’air française a l’entrée en guerre,” Rev. Hist. Deux. Guerre Mond. 19 (January, 1969): 111–16.)

 As a professional, General Truelle finds the question of how many more aircraft can sortie on a given day than global totals, and prefers long production series to bullet points in order to understand an industry's overall trajectory. French a/c production, he notes, is 01/37 32; 10/37: 99, 01/38: 41, 08/38: 77, 10/39: 357, 02/40: 188, 06/40: 684. Marked seasonality shows that the industry had not overcome exogenous drags to production, but a highly favourable structure weight per person output shows that the industry's problems are balanced by advantages. Citroen Nation could make stuff.

And notice that huge fleet of single engine fighters? 399 "transitional" types and 240 modern ones? It doesn't even include the contract for 1000 Morane-Saulnier M.S. 406s. Deliveries will be 11 a/c per day by the outbreak of war, and over 500 will be in service, so presumably Danel's figures exclude a large number of aircraft not formally cleared for procurement, a not uncommon strategy sometimes used for various reasons to minimise the number of aircraft in a given military service. . (600 Swordfish and 200 Sharks on hand versus a Fleet Air Arm of supposedly only 150 a/c on 3/09/39 *cough* *cough*). The full Morane-Saulnier contract will complete delivery by March of 1940.

Compared to the Hawker Hurricane, we can certainly see a case for criticising French industry here. Like the 406, the Hurricane was a transitional structure design begun in 1934. The thought behind this was that although the Hurricane would be unnecessarily slow due to the use of old-fashioned materials, it would come into service more quickly than a more radical alternative. The same argument could be, and I believe was, pressed with the 406. So the fact that the Hurricane made it into service almost a year earlier than the 406 is a comment on the relative state of the two industries.

Except that at the outbreak of war, there were 500 Hurricanes in service, same as with the 406. Hare, tortoise, like that. The key difference is that the Hurricane production contract was extended. More to the point, there were 549 Hurricanes in squadron holdings on 15 August 1940 and another 189 in depot and service holdings. This is impressively more than the famous "610" fighters supposedly held by Fighter Command "during the Battle of Britain" and impressively less than 1000. Accounting for attrition, it is possible to understand why aircraft contracts were so much larger than in-service totals, and thus why production contracts at the outbreak of war were so large. The fact remains that as far as fighting the Battle of France went, the men and women behind the Morane-Saulnier M.S. 406 had done their job. There were enough of them. If France needed thousands more single-engined fighters, it could look to other builders.

And it did, and there were. Dewoitine was bringing in its famous D.W .520,  successor to Plan III's cantilever monoplane 500 series. The 520 contract was placed in 1936, making this aeroplane a year younger than the Spitfire, although again it turned out a year later than expected, putting it in the class of 1940 along with the Beaufighter, Grumman F4F Wildcat, Nakajima Ki-43 Hayabusa and Mitsubishi A6M.* Fourteen hundred D 520s were on order at the outbreak of war, with deliveries to peak in May at 200/month. Combat clearance usually lags first production type flight (October 1939), so it is not surprising that les avions de chasse Dewoitine DW 520 were not cleared for action until April 1940, and if you think that I lapsed into French there just for an excuse to say "avions de chasse," you're completely right. The chivalry of France was clear to ride. (La Mireille.) 228 520s had been received by the French air force, and 75 accepted. Only Groupe de Chasse I/3 was fully equipped with 36 types, but four more groupes and 3 naval escadrilles were re-equipped during the battle.

This actually isn't that terrible. The first Spitfires entered operational service in August, 1938, but as late as 15 August 1940, only 276+207 were available. No Spitfires, rather notoriously, fought in the Battle of France at all. 

Nor is this story complete. Contemporary to the 406 was the Bloch M.B. 151, a bit of an orphan on Wikipedia in spite of being the main French fighter of the battle, equipping nine full Groupes de chasse. The best of the old splatbooks to hand suggests that the full contract for 500 aircraft had been long since completed by the outbreak of the Battle, but limply suggests that the well-known problems with gunsights and propellers had kept the aircraft grounded, leaving the pilots without experience of this potentially highly effective aircraft.

Okay, but no. Enough with the apologies here: you could defend this claim by pointing out that the Gnome-Rhône 14R engine that powered the Bloch MB150 series was comparable to the Grumman F4F's P.&. W. R-1830, or the various Mitsubishi engines powering the A-6M Zero, and since those planes were great, there wasn't a problem with the Bloch that couldn't have been fixed. And then I'd have to stamp on your special snowflake Pacific War-walled garden aeroplane and it wouldn't be pretty, and that would be awful.

No, wait. On second thoughts, that sounds like awesome fun.

So we've got two issues here: the first is that the French didn't build enough aircraft. And this is a pretty hard argument to make. Not enough bombers for offensive operations, sure. Not enough fighters for defensive operation, now that is a hard case to make. So if air war is all about defence with fighters and ground support operations with obsolescent fighters (hint: it's not), then the French were perfectly well set up. 

Still, the reason for not having enough fighters specifically to break up the carpet bombing of the Sedan position, the second argument, that they didn't have enough good aeroplanes at the front on 13/05/40, is pretty easy to argue, on account of being, well, true. 

The question is what made the (theoretically enough) planes into a critically deficient agglomeration of quality? And the answer is, at the end of the day, not the planes, but the engines. Pilots can't make bricks without straw, and the Hispano-Suiza 12Y and the Gnome-Rhône 14 were straw. And this is kind of a problem. One of the great errors of the old production history of World War II was that aircraft were somehow like automobiles, and that the Allies' great triumph was to turn out aeroplanes like Model Ts, using the same techniques. David Hounshell blows that myth up so good that he should change his name to Michael Bay.  That being said, while airframe production is very much a specialist business, aeroengine production is not. It is very much in the mainstream tradition of making power plants for other purposes. The firms that flourished in aeroengine manufacture during WWII, either building on their traditional niche or invading a new one, typically made either automotive engines or electrical generators, or both, before the war. Ironically, it was precisely and only the American specialist aeroengine makers that ought to have prospered from all that "assembly line" production during WWII, that failed to make a successful transition into the postwar era.

So what happened in France? Hispano-Suiza was a well known automotive engine builder, and while Gnome-Rhone specialised in aeroengines with only a minor sideline in motorcycles, it was a big builder. One answer is lack of public support: F. R. Banks once estimated the cost of developing a new engine at an established firm w. research branch, etc, as:  prototyping, at £55,000, plus, for 10 development engines to be run for 5,000--10,000 hrs type-testing , £1 million –2 million. . ("Letters," Flight, 24 March, 1949, 358.) At the same time, both of the major French fighter engines had been developed. Leaving the G-R design aside, why did the 36 litre Hispano-Suiza deliver 20% less power than such other 27.1 litre Merlin. (Or, to make the comparison more global, and switch to Aviation's preferred cubic inches, 2070 for the Mercedes Benz 601, 1710 for the Allison, 1647 for the Merlin X, and 2197 for the Hispano?)**

It's certainly not weight; the Hispano is the lightest engine on this list by a significant margin. We can also see, or rather, German engineer C. Michaelis can see, by directly comparing captured examples, that both French engines are quite weight efficient, while American examples are actually quite inefficient. Indeed, I'm not going to pursue this comparison at length, but it's pretty obvious why Wright failed and Pratt & Whitney let GE and Westinghouse into their business, and rather less obvious  how the American air forces actually managed to win the war.*** 

Now, the Wikipedia article attributes the problem to inadequate aspiration leading to a lower engine speed. But there's more to it than that. In automotive engineering, large displacement is traditionally an alternative to  higher engine speed. A lower bmep/inch is compensated by having a larger piston area. If you don't buy the additional swept volume with excessive weight, as Hispano certainly didn't, you have an engine that not only potentially delivers as much power, but does so with lower maintenance costs. That being said, it is certainly not what Hispano-Suiza achieved. The reason for this is that, while being quite metallurgically sophisticated, the Hispano engine is designed to be cheap. A particularly outrageous example of this is the use of case-hardening steel that is not hardened. And so on through the design, one finds often quite sophisticated design in metallurgical terms, for example the widespread use of nitriding by the firm that, after all, pioneered it, but in the service of unit cheapness and low maintenance costs rather than performance. It is weird in this light that the French should have made so much use of exotic alloying material. On the other hand, at a time of global trade stringency in general, the French had no trouble importing things. 

What I have not so far said about the French fighters of 1940 is that they were typically much lighter than their rivals or contemporaries. The 151 had an empty weight of 2158kg, compared with the Hurricanes 2,605 or the Wildcat's 2674. The 520 weighed 2,123kg, and the M.S. 406 came in well under 2000! This comparison has sometimes been drawn by professionals seeking to discredit the ever-green illusion of the "light" fighter.

The main French military aeroengines of 1940 were, in detail, good engines. It would have been an easy matter to redesign them to give considerably more power at considerably greater expense, and this was, in fact, under way. The aircraft could probably not have taken bigger engines: the Dewoitine in particular had vicious ground handling. But, then, the whole point of letting numerous small-scale production contracts to different design firms is to "fail forward" in small production branches, each new aircraft bringing new possibilities, and, of course, new problems. Arguably, the French had to stop doing that if they wanted to meet really ambitious production goals. On the other hand, it is hard to see what long term advantage Britain gained from producing 14,000 Hurricanes, or Germany from almost 20,000 Bf109Gs. 

To put it another way, you can take the automotive engineers out of the automotive business, but you can't get rid of the automotive engineering quite so easily. Told to build an avion de chasse, the French went forward with a sports car, complete with the cheap, mass-produced version of the racing engine that they would put in their "speed" variant, when perhaps what they needed more was to design a speed locomotive. And then plan on building 10,000 of them.

So the basic answer here, it seems to me, is that air forces need to spend excessive amounts of money on ambitious new technologies that don't actually work that well. And when I say "excessive," you really have to lose your conservative habits. It may seem like no-one is going to buy a sports car that costs as much as a sports car would cost if they didn't fall apart when you drive them around the block, but this is government work. Oh, sure, when you need to chuck in your hand and mass produce whatever  happens to be in the queue, starting a few years before the beginning of the war, people are going to whine about it not being designed for mass production. Who cares? That's what other engineers are for.

The next question is, where did the money go? My provisional answer is that this problem is akin to finding the leaks in a hose. All you can do is pump water in and see where it comes out. My impressionistic answer is that while the French pumped a lot of money/water into the pump in 1938/9 without seeing planes come out the other end, they did, in good time, see curious and impressive little cars that still provoke enormous customer loyalty today. As I said to start: once the government may start with the intention of spending money on stuff with a very clear agenda of preventing involuntary auto-regime change, but once the money  has left the Treasury, it is likely to find its way to unexpected places. 

* Just because this is fascinating generally.

Gladiator, Hurricane, Polikarpov I-16
Bf109E Fiat C.R. 42, Bf110, Macchi MC200, Nakajima Ki-27
Beaufighter, Dewotine D.520, Grumman F4F, Nakajima Ki-43 Hayabusa, Mitsubishi A6M
FW190, Macchi MC202, P-39, P-40, MiG-3, LaGG-3

P-38 Kawasaki Ki-45 Toryu, Yakovlev-1
Macchi MC205, Fiat G.55, Me410, He219, Kaw Ki-61, Nak Ki-84 and Ki–44, Grumman F6F, Vought F4U, P-51, Lavochkin La-5, Yak-9
Tempest, Me-163, Me-262, P-61, La-7, Yak-3
FW Ta-152 (FW190 follow on, not “Moskito”), Meteor, Kawasaki Ki-100

From the dishearteningly obscure Angelucci, Enzo and Paolo Matricardi. Hunter of the Sky: A Visual Guide to World War II Aircraft (Abridged ed. New York: GT Merchandising and Licensing, n.d.; original edition Milan: A. Mondavi, 1990). This is a splat book with insight. Is it patronising to say of it, as of the best Italian scholarship, that it is thorough and interesting and comes at things from a completely novel perspective?
**Raymond W. Young, "The Mercedes-Benz DB-601A Engine: The Results of an American Examination of a Specimen Captured by the Allies," Aviation, October, 1941, 270--9; 300--01.)
***P. Kӧtzschke, "A German View on Allied Engines: An Enemy Report on Five English, Three French and Two American Types Captured," Aircraft Engineering, September 1941, 240--6.


  1. OK, so the French had a lot of aircraft and some of them were pretty neat, and the production flow was quite impressive as well. I think we knew that.

    So what happened?

    Strategically, they were OK (they built a lot of aeroplanes and deployed the great majority up north in the battle area).

    Operational art: ? Was it that they weren't concentrated where they were needed and when, within the ZOAN?

    Tactics: ? Was it that for some reason they tended to lose section-level dogfights?

    Logistics: ? Was it that they couldn't be adequately supported, and therefore couldn't keep up the sortie rate?

    We know the French had some sort-of radar, and the RAF brought over quite a bit of mobile radar in 80 Wing, but it never really worked.

    So, was it Colonel Mustard in the billiard room with the lack of a telephone? (Note that the collaborators chose Vichy as their capital because it was the one city outside Paris whose local telephone exchange could make international calls.)

  2. I'm going to provisionally say that the French problem was, first, that they didn't have an adequate counterforce (not enough heavy bombers). Second, it's what everyone says. Their fighters weren't fast enough. They could have been fast enough with engines that were designed to be re-engineered for higher performance at higher cost.

    That didn't happen, and, I'd add, would probably have made them pretty dangerous planes.

    What were the consequences of this? Well, first, they problems in air combat. In particular, most shootdowns happen by surprise, and so the key fighter ability is to escape from a disadvantaged position. Sometimes, you can use superior handling to do that, but just plain being faster is far more important.

    So, that said, let's look at these planes. Sure, we have published Vmax, but these range from being a bit misleading to being outright fraudulent. So I'll look at power loading based on empty weight. (Which can be problematic, too, but you can only go so far into epistemic crisis!)

    Bf109E:lb/hp (empty weight): 4.01
    Bloch MB-151: 4.40
    Hawker Hurricane:4.53
    Morane-Saulnier MS 406: 4.87
    Dewoitine D. 520: 5.08
    And for comparison's sake: Grumman F4F-4 Wildcat:4.91

    That the 520 had an inferior thrust-to-weight to the 406 and especially the MB-151 is an illustration of the importance of aerodynamics to the Vmax figure. However, viscous drag is proportional to the square of the speed. The slower you're going, the less important it is.

    What that means, on the one hand, is that acceleration out of Vcruise (and thus out of disadvantage) is going to be higher for a plane with a higher thrust-to-weight ratio, all other things being equal.

    It also means that the sweet spot of efficient cruise is going to be higher for the plane with the higher thrust-to-weight ratio. Vcruise, and V(max safe engine-rating power) are rarely published for WWII fighters, but they are the operationally crucial speeds, along with climb rate, which is a function of wing-loading, which is also very intermittently published in splat books.

    Given the low weight of the French fighters, I can say with some small confidence that their wing loading is likely to be high --very high in the case of the D 520, as there is no other way of getting that speed out of it.

    The French planes will have difficulty getting to the fight, and in capitalising on initial advantages in air combat. They will have trouble getting out of disadvantaged positions, and they will suffer a relatively high rate of ground accidents.

    Not that the Bf109E is the queen of the landing strip, either!